Pending fault codes P1171 and P1174, tuning advice needed please

[Bone9]

The banks do vary a bit yes. Before they were very similar.

[LPGC]

Fuelling between banks would vary a lot if (say) an injector on one bank wasn't working at all (no fuel to one cylinder). But in that case the full extent of the difference in fuelling could be masked if, say, fuel trims on both banks were high positive before the problem. For example if you have trims at +15% and the maximum they can reach is +20% then you'd only see a max difference of 20-15=5% in fuel trims even though one cylinder bank would be 25% lean on average because one of the 4 cylinders on that bank was getting no fuel at all. If instead of that last example you originally had both bank fuel trims at zero% then no fuelling on one cylinder would see the full 25% difference reflected in fuel trims (at least before the system noticed the problem and switched to open loop fuelling). Of course if a cylinder got no fuel at all (or even lean or rich by more than around 20%) there would be a misfire on the affected cylinder anyway.

The above probably won't help you much so I'll post some additional thoughts...

When lambda sensors are failing they are more likely to still work correctly when running on petrol than they are when running on LPG.

When it was working well before you still had injectors that didn't really flow enough, the flow rates of injectors was too small because the internal size of the V34's is relatively small (so you couldn't correct it by fitting bigger nozzles), your gas injectors duration still had to be way higher than would-be petrol injector duration for the engine to see correct fuelling. Such a setup is flawed from the outset (with some ECUs more than others) because the default ECU temperature correction factors presume gas injectors opening/closing time and the multiplier you calibrate (petrol injection pulse length to gas injector pulse length) should normally see gas injectors pulsing at (say) 1.1 x the duration of petrol injectors - Having gas injectors pulse for much longer/shorter than the example 1.1x can mean the temp correction throws fuelling off. Every time there is a change in temperature of the weather (seasonal changes) I fix a lot of LPG conversions where the installer didn't use appropriate injectors / nozzle size / pressure / temperature correction. AEB based systems compensate fuelling not only for gas vapour temperature (the only temp that really matters because it's the only one that affects the density of the gas vapour at a given pressure) but they also compensate gas fuelling for reducer temperature. Reducer temperature is something that shouldn't really matter, only gas vapour temp matters. Seasonal changes can mean the car switches to LPG sooner, but in doing so the engine hasn't been running for as long and this can in fact see the reducer temp fall lower than it would if it switched to LPG at the same temp in winter (when the reducer started at a cooler temp and took longer to work it's way up to the switchover temp).

[Bone9]

Hi Simon sorry for the late replay. Interesting information about the fuel trims.

How can I test to see if the lambas are functioning OK?

The software says my injectors are one line over half, towards saying the injectors are too large. I did wonder about the temperature difference and thought that might change the trims. But wouldn't the trims be richer when it's cold as the air is colder and more dense? When my map was last edited (February) and running great the weather was much cooler. Please see the attached pics for weather info. I nearly bought a new set of Valtek 34's. What I injectors would you recommend for my setup?

Please see the temperature correction setting also.

Thanks.

Toby.

[Bone9]

What do you think @LPGC?

[LPGC]

Unless the installer wired the LPG ECU to the lambda sensors there isn't a straightforward way of knowing if lambdas are working properly from just looking at the LPG software screen... There are subtle ways someone with experience can tell by but they're nothing like as easy as watching the lambda voltages 'flick' while you watch them. You could plug an OBD reader that shows live data into the OBD port and monitor lambda readings to see if they flick, that would be the same as watching lambda readings flick in LPG software if the LPG ECU were wired up to lambdas. If there are no misfires and pinj readings on both banks are fairly equal then that too points to lambdas working OK because we wouldn't expect both lambdas to fail at the same time while if one lambda worked OK and the other didn't we'd expect to see pinj readings different between the 2 cylinder banks.

The software's advice on injector nozzles being too large/small should be taken with a massive pinch of salt, it isn't very accurate or useful for a decent installer but I suppose it could be used as a check for someone new to setting up LPG systems. It probably worked a bit better with older types of LPG injectors than it does with modern injectors. The way it works is it expects ginj to be (say) 1.3x longer than pinj during autocalibration... while autocalibration on older AEB systems involves holding the engine at say 2500rpm with no load, so it's expecting the 1.3x under low load conditions. Behind the scenes the expectation would still be that 1.3x multiplier would come down to (say) 1 or 1.1x at higher loads when the offset of the slower response of LPG injectors is a lower overall percentage of total pulse duration. Older injectors were slower to open than modern injectors and much of that example 1.3x *(the extra 30% of pulse duration compared to petrol injectors) would be assuming that slower response of the injectors. If the ECU gets the programmer's assumption of injector response time and/or flow rate when it is partially open wrong they cannot get the 'injectors too small / injectors too large' part of the software to work accurately... a case of bad information in bad information out, especially if we're talking idle pinj that isn't the usual (anticipated) 3ms. We wouldn't want ginj to be pinj x 1.3 (pinj +30%) at higher engine loads because there comes a point when pinj is just short enough to avoid running full duty (not getting time to close if they're going to have time to close and open again in time for the next injection pulse). I've mentioned full duty and 'injection window' (which is time period between 4 stroke cycles) on forum many times before. Window in milliseconds is 120000 divided by RPM (e.g. 6000rpm), so window at 6000rpm would be 12000/6000=20(milliseconds) and we usually find that maximum pinj isn't far off window at the engine's red line.. unless the engine makes a lot more torque at lower rpm than it does at higher rpm (in which case that pushes window at the red line up, in which case we'd normally find peak pinj is also up) or vice/versa.

If the gas vapour temperature reading is accurate there should be little to no need for the reducer housing temperature reading. There's really only AEB systems that use reducer housing temperature reading as an input that effects fuelling by default (there are many systems that can be set up to use reducer housing temp reading to adjust fuelling as an option). Most modern systems use a combined gas pressure sensor and gas temperature sensor that's fitted in the main feed line between reducer and injectors, these tend to respond quite quickly to changes in gas vapour temperature. Back in the days of some of the older AEB systems the gas temp sensor might have been fitted at the far end of a gas injector rail, which would never really have any gas flowing past it (because it wasn't in the line but was at the end of the line). Those older gas vapour temp sensors were more likely to pick up the heat of the injector rail itself, which would probably be close to general under bonnet temperature, than to accurately measure the temperature of the gas vapour flowing to/through injectors. So the reducer temperature could be useful as an additional pointer to correct fuelling (correct ginj) for changes in temperature of the LPG flowing through gas injectors. We shouldn't really need to correct for air temperature because MAF sensors and/or IAT sensors already correct engnie intake air mass for that and automatically correct pinj for the temperature difference in density... But we do have to correct ginj for gas vapour temperature because both the pressure and the temperature of gasses (including LPG) affect it's density. There is no MAF sensor for LPG but we do have a pressure reading and we do have a temperature reading. These days with a modern system if
enabled reducer temp correction at all I'd be more likely to dial in a positive reading for cold reducer temperatures than a negative one because cold injectors can be slower to open than warm injectors, the cold reducer correction would normally only apply for a short time after changeover to LPG and the gas vapour temp correction is quick to respond and accurate constantly under almost all conditions.

What happened when you turned off one injector at a time?

[Bone9]

Thanks Simon

Right OK. My o2 sensors seem to be working correctly as I can see them flick between volts via my torque pro screen.
The attached photo shows my current petrol trims.

At idle my ginj is about 1.4 times my pinj at idle.

I also noticed my injectors hoses vary in length between 160mm and 200mm.

I do believe my gas temperature sensor is in the end of the fuel rail but I'll have to check tomorrow to confirm.

For my idle to have the same pinj I had to increase the box from 145 to 165.

When I turned each injector off I didn't see a difference. This was done at idle. I did try while driving but needed a better helper.

I'm wondering if It's just my map that needs changing.

[Bone9]

The gas temp sensor is indeed in the end of the rail.

[Bone9]

Hi Simon. I've also just checked the size of my nozzles and they are only 2mm. I was thinking of opening them up to 2.5mm.

[LPGC]

Yes I would open the nozzles to 2.5mm but don't forget you'll need to recalibrate if you change nozzle size and/or pressure.

[Bone9]

OK will do. Do you think my map was probably never 100% and that why it's out now?

My reducer does still output a much higher gas pressure when running on petrol. I have a gasket rebuild kit here. Is it worth stripping it down and rebuilding it?

[LPGC]

If the combination of nozzle sizes and pressure isn't well suited then you'll struggle to map it well anyway, plus the temp correction won't work as well as accurately as it should. But whether or not you made the best job of mapping (even considering the above mentioned compromises) may be a different question.

Would really need to watch the pressure to see what's happening when you switch to petrol to know for sure if there's a reducer problem. But such a high pressure reading (you're seeing over 3bar?) from a reducer that's only outputting 1.3 to 1.7bar when running on LPG does point to a reducer issue. It's not the type of issue that replacing a diaphragm would fix, I'd expect such issue to be due to a faulty inlet/metering valve inside the reducer... Not a diaphragm or solenoid problem, could be a gasket issue but I very much doubt it. If rebuilding the reducer (without changing hard parts such as the inlet/metering valve) fixed it, I'd expect the fix to be due to disturbing the connection between the diaphragm and inlet/metering nozzle. You don't normally get inlet/metering nozzles in rebuild kits.

[Bone9]

OK Simon, I'll drill the jets to 2.5mm tomorrow and make a vid of the gas pressure at the weekend. I won't bother rebuilding the reducer. I'll replace it if you think it's faulty. My lpg pressure when on gas is 1.6bar but over 3bar on petrol. Will I need to reduce the lpg running pressure for the new jets?

[Bone9]

Hi Simon. The jets are now 2.5mm and running better but over fueling. My gas pressure was ok at the weekend so couldn't get a vid.

Do you have any availability on the 22nd of this month? I'll drive the 3.5hrs each way for you to map and sort out my lpg finally.

Thanks.

[LPGC]

Hi Toby,

Sorry for the late reply.

I replied to your Whatsapp message. Yes I can see you on the 22nd.

Simon

[Bone9]

Hi Simon. Great, thanks. See you on the 22nd.

[LPGC]

Hi Simon. Great, thanks. See you on the 22nd.

OK, see you then.

[Bone9]

Hi Simon. Thanks for the mapping. Its been great on the longer drives and have done 2 full tanks with no problem but I've got an issue with my long term fuel trims that seem to be from when drive on short journeys from cold. I only travel 4 miles to work each way. This sends the longterm fuel trims to -14 or more on both banks. This will settle back down but only after longer runs. Problem is most of my journeys are short so it doesn't get a chance to sort itself out.

[LPGC]

Hi Toby,

You're welcome.

Trims hitting -14% on short drives is not necessarily an issue. Have you compared to trims when running on petrol in the same (4 miles run from cold start) conditions?

If trims are very different between petrol and LPG under those driving conditions they can probably be made closer by adjusting temperature correction parameters in software. But because of the way your system is designed (type and location of vapour temp sensor) the temperature reading won't quickly respond to actual changing temperature conditions. With a slowly responding vapour temp sensor and/or a vapour temp sensor design that is prone to being affected by under bonnet temperatures (not just responding to gas vapour temp sensor) there'll always be a bit of a difference between cold trims and warm trims because the extent of the lag (response time to actual changing temperatures) can't be compensated for because the extent of misreadings due to lag will change with the conditions.

For example, you start the engine on a cool (for this time of year) morning, everything might be at (say) 10C. It switches from petrol to LPG when the reducer/body/water temp reading is 30C at which point the vapour temp reading will still be 10C. If the engine is driven at light load at that point (so there isn't much throughput of gas through the reducer, so the gas flowing through the reducer spends a relatively long time insider the reducer where it is heated by the body of the reducer) will for a very short time be 30C quickly falling to (say) 20C... But the sensor doesn't immediately reflect the 30C or the 20C, instead it takes time for the sensor to warm up (lag). Now let's suppose after 3 minutes driving on LPG the reducer body temp is 40C, light throttle load so the gas output from the reducer is at 25C, the sensor still reads 20C.. But you put your foot down so now the gas coming out of the reducer is 15C but the sensor still reads 20C. You drive the car for longer still then get stuck in a traffic jam, the engine bay temp increases and because of the design of the sensor it starts to pick up under bonnet heat instead of gas heat. If you've got a sensor that has too much lag or is inaccurate (lag implies inaccurate anyway) it can be better to dial down it's bearing on (or how it affects) mixture, because responding to inaccuracies in some conditions can be worse than responding to inaccuracies in other conditions...

[Bone9]

Hi Simon. I haven't tested the cold Petrol trims yet but LPG LTFTs hit 16.5 on the way home from work.

On the way to work this morning I got an engine management light and these codes.

[LPGC]

Was the high LTFT just on one bank or both?

If you tried switching to petrol during the time of the high fuel trims, what were the trims on petrol?

I know you rebuilt one injector that failed, it could be that more injectors are failing.

Also they're a type I wouldn't have advised/fitted for the application and they're plumbed into the plenum rather than into the inlet manifold.